World Cities in a World System*


* Article publié dans Hérodote, n°101, 2è trim. 2001. Translated by Sean Duffy.


From the region close to Birmingham – a city as renowned as its cousins Manchester and Liverpool during the great age of the British Empire and which has suffered, like they have, the effects of the shifting of the axis of the world system in the twentieth century – comes now a stimulating study of world cities i.e. those cities which play a driving role in the present world system. This research is led by the Globalisation and World Cities Research Group (GaWC) which publishes its work on the Internet as Bulletins.1 The Department of Geography at the University of Loughborough – 150km to the north-east of Birmingham – is the linchpin of this research which also has input from other universities and disciplines. Amongst the leaders of this research project, Professors J.V.Beaverstock, D.R.F. Walker and P.J.Taylor play a notable role. I shall for purposes of simplification make exclusive reference to the latter, in addition this will allow me to focus on his works which are – in my opinion – of the most significance, as the reader will be able to judge further on.2


The world city network

Taylor and his team take as their starting point the observations of Castells on the advanced services from which societies in the process of rapid computerisation draw a large share of their resources.3 By locating the cities where these services come from, by observing the permanent international links that their productive businesses maintain from city to city – with the help of subsidiary companies or alliances, by deciphering the diverse variants of the network formed in this way, GaWC throws some light on the functioning of cities which irrigate the world system by offering the most advanced services to the centres of production and exchange.

The services referred to have been selected following a succession of approaches. In the first instance, the press of major cities was considered and the management of large businesses. Daily newspapers of these major cities which abound with business information were analysed to judge about which cities their local readership were informed. In addition, the role of highly-qualified migrants from these chosen cities was also taken into account – although they are difficult to track – the specialised press and enquiries with personnel management of international firms being the only sources available – access being difficult regarding the latter.

In contrast, some international services of first-rank importance have been deliberately ignored, when, as in the case of air links, it is difficult to discriminate between everyday consumption such as tourism or the travels of individuals, and the benefits of services which are indispensable to businesses with a worldwide vocation. The above approaches have been judged to be too qualitative both to provide an image of world cities which can be periodically measured and to determine the networks that these cities form.

Consequently, attention is brought to bear on four categories of services of which multinational firms are the significant users and whose production is guaranteed by businesses which have themselves spread their networks of subsidiaries and participation to all the cities where their customers wish to find them. Accounting (and auditing), advertising, finance (banking included), and insurance have thus become central reference points, it being understood that in these diverse branches, commercial, tax and financial law are fully represented. Within these diverse categories specialised knowledge, useful to customers, can be adapted to legal, fiscal and even the most varied type of cultural context – which is precisely one of the essential demands of businesses in the course of globalisation.

Forty-six business producers of advanced services introduced during 1997-1998 in fifty-five world cities have in the end been retained as a representative , starting out from an initial selection of businesses in 142 cities from an initial 263. The whittling-down process has been operated by eliminating businesses not making use of agencies in at least 15 cities, which limits the matrix of potential connections to 46 (businesses) in 55 cities, that is to say 2530 locations. It is due to these that the networks managed by the producers of advanced services and the networks of the interconnections of the cities can be measured with calculations justified by statistical theory. While not going into these investigations – detailed by GaWC Bulletins , I will note that the services considered are obviously heterogeneous to the extent that taken together they do not reveal a network of formal coherence which would be comparable to that of a national or international economic accounting.

Nevertheless, this is a similar area to that where the Index of human development4 published by the UN Programme for Development subjects the Index of National GDP5 to a very rigorous scrutiny. Allowing for a few minor variations, GaWC has operated in a similar fashion, whilst detailing stage by stage in its Bulletins the methodology it has used, it has not concealed its hesitations or occasional regrets. Its results are therefore based on a reasoned approach, as well as on the available facts, and on the statistical techniques that have permitted its activities. The total corpus of their research – provisional, because the research is ongoing – permits one to typify the diverse network of advanced service providers and to hierarchize, more or less, the cities with a worldwide vocation which were observed in 1997.

As regards service providers, GaWC details diverse models. Thus, the strategy of American law firms is to maintain a lobby in Washington DC and a financial antenna in New York, whilst the London subsidiary is the centre for international activities. In contrast, British law firms focus on European cities and on the financial areas of Tokyo and Singapore – agencies are rarely found in the USA. A final example; advertising firms, take an interest in many sectors of the European market but rarely compete with their American and Japanese counterparts on their own turf.

As regards the hierarchy of world cities, it is ordered according to a simple criteria. The producers of the four branches of the services previously cited are arranged into three classes according to their global size. The leading class is given a value of 3, thus a city where the four main branches are represented by businesses of this class gains a score of 12 e.g. Paris. Similarly, a city in which there are first class businesses in two of the branches and two other branches containing service providers of average class receive in total a score of 10, which is, in another example, the case for Frankfurt.

This classification continues on down to the cities where the four branches of services are only represented by businesses belonging to the final class (worth 1) which gives them a score of 4, as is the case in 1997-1998 – as I remember – for Berlin, Hamburg and Shanghai. Below a score of 4, cities are not considered to have reached a sufficient level of world influence, which, amongst others, is the case for Birmingham and Rotterdam.

A - ALPHA WORLD CITIES (full service world cities)

12: London, New York, Paris, Tokyo

10: Chicago, Frankfurt, Hong Kong, Los Angeles, Milan, Singapore

B - BETA WORLD CITIES (major world cities)

9: San Francisco, Sydney, Toronto, Zurich

8: Brussels, Madrid, Mexico City, Sao Paulo

7: Moscow, Seoul

C - GAMMA WORLD CITIES (minor world cities)

6: Amsterdam, Boston, Caracas, Dallas, Düsseldorf, Geneva, Houston, Jakarta, Johannesburg, Melbourne, Osaka, Prague, Santiago, Taipei, Washington

5: Bangkok, Beijing, Montreal, Rome, Stockholm, Warsaw

4: Atlanta, Barcelona, Berlin, Budapest, Buenos Aires, Copenhagen, Hamburg, Istanbul, Kuala Lumpur, Manila, Miami, Minneapolis, Munich, Shanghai

The world cities are divided into three groups according to their score. The alpha group scores between 12 and 10, whilst in the gamma group, the final group, scores range between 6 and 4. Thus, those world cities which truly have a world role; London, New York, Paris and Tokyo, appear in the alpha group, followed closely by Chicago, Frankfurt, Hong Kong, Los Angeles, Milan and Singapore. The following category or beta group – designated as containing major world cities – opens with San Francisco, Sydney, Toronto and Zurich (score9), followed by Brussels, Madrid, Mexico City and Sao Paolo (score8) with Moscow and Seoul closing the category with a score of 7.

Next comes the first section of the gamma group (those with a score of 6): Amsterdam, Boston, Caracas, Dallas, Dusseldorf, Geneva, Houston, Jakarta, Johannesburg, Melbourne, Osaka, Prague, Santiago, Taipei, and in an irony of alphabetical order - Washington.6 That Washington is in this list – making it less of a world city than Moscow – and that Rotterdam is not ranked amongst cities which already have world status should not lead one to a hasty conclusion, for it is not a question of intrinsic defects in the survey, rather that it has clear limits. It is better to judge it by its benefits, before wondering how one can overcome its limitations.

Cities with global influence, identified as those with measurable advanced services, up to the present time, only feature in one state out of six. Obviously there is a greater density of them in North America and Europe than elsewhere, even than in East/South East Asia where they are spread out from Tokyo to Singapore.

Their classification establishes a clear hierarchy, as well as regional configurations. Take Miami as an example. Its position hardly corresponds to its role in the North American subsystem, rather it results from its financial pre-eminence with regard to Latin America. In contrast, the exclusion of Rotterdam, one of the major ports in the world and a major port of entry for the oil consumed in Europe, shows that this important commercial city is relatively underdeveloped as regards the services firms look for in order to develop on a worldwide scale. Detailed studies on a city or subsystem basis abound in the GaWC Bulletins alongside works concerning the overall world network.


The information gathered in this way is of great value and will become even more so as investigations carried out at regular intervals will permit one to observe the transformations of the networks of world cities and how business clients use their services. However, these investigation will encounter practical difficulties which will only be overcome by the use of skilful methodological precautions.

The sample of firms observed will be of necessity disturbed by the appearance of new businesses, as well as by the disappearance of companies which had previously been considered (due to decline, absorption by another group or the break-up into several groups). Another risk will be of the elimination of one of the four branches of activity retained by the 1997-1998 study, for example by an increasing interpenetration of banking and insurance activities. On the other hand, one might hope that new categories of firms with a worldwide vocation will broaden the field of study towards services which help in the location and de-location of firms (marketing, ‘political’ risk evaluation etc) or who approach more closely still the strategies of multinational enterprises. Disturbances in the examples to be identified will change the significance from one enquiry to the other, whenever they are held, from ‘who does what in terms of services on a worldwide scale?’ towards ‘how does the geography of their world network evolve?’


Already it appears that the world network of cities resembles national networks which have been studied for a long time: it is one of the essential reference points of the multiple transformations which affect the area where it (the network) is spread out. In contrast to previous decades where the economic movement of globalisation appeared to be transmitted from one state to another, then as a process of internationalisation affecting one branch of the economy after the other (for example the chemical industry following the automobile,etc) it appears henceforth as the work of multinational firms, fed in particular by the network of world cities and whose choices are changing, varied even capricious. Again it would be desirable to extend the enquiry beyond the advanced services to include the firms served by them, notwithstanding the confidentiality which is held dear by business. In effect, the direct exploration of the uncharted continent formed by the multinational firms themselves is an objective which university research ought to try and base on the co-operation of states and certain international agencies, such as the UN Conference for Commerce and Development. This vast group of businesses on the move, hungry for stock-exchange and commercial publicity, jealously guarding its banking secrets and its use of havens (for tax or other reasons) is to-day the main motor of the world economy, the mechanism through which the mode of capitalist production is spread. The advanced services do permit one, admittedly, to get close to these firms, but only to a certain extent and from a particular angle. They reveal nothing of inter-banking operations and their stock-exchange repercussions, thanks to which the capital flow spreads out or falls back, on the whim of multinational groups, dependent on their technical and market perspectives. They reveal nothing either about the criteria (market, salary, tax, security, etc) which determine the geographic preferences of these firms. Nothing, in the end of the way in which their resources, waiting for investments of fixed capital, are mixed with capital which floats from one country to the other, according to the fluctuations of currencies and monetary policies.

The services looked at constitute the judicial, tax, commercial and financial circle around the multinational firms, but they are not their very essence. The cities where these services bring their collaborations together are admittedly not a simple conjunctive tissue which would surround the vital organs of capitalism. They are for the latter a valuable auxiliary motor, but nothing more. As the example of Rotterdam underlines, it is from other cities of worldwide importance where commodity capital flows strongly.

On a more general level, the network which generates the international flow of capital will be increasingly understood as one gets nearer the heart of the capitalist system: the strategies of multinational businesses, including those of the banks that serve them and of those of the most innovative or profitable firms, these latter deriving their profits from products which will become increasingly scarce – such as oil, or in the future, water.

see : world cities n° 2


1 -The address of GaWC is

2 -. A great number of bibliographies - too often limited to the Anglo-Saxon world - appear alongside the research published in the Bulletin of GaWC. Those authors who very often re-appear are: M.Castells The Rise of the Network Society, Oxford, Blackwell, 1996, edition francaise : La societe en reseaux, 3 vol. Fayard, 1998; J. Friedmann, The world city hypothesis, Development and Change, no 17 (1995) and S.Sassen The Global City, Princeton University Press (1991) and Cities in a World Economy, Pine Forge Press, London 1994.

3. Op Cit Volume I Chapter VI.

4. The justification of such methods of collection is given into in great detail by Amartya Sen, in On Economic Inequality (Oxford and New York, 1973) and in Inequality Revisited (Oxford, 1992).

5. The GDP or Gross National Product assesses the net annual production of a given country.

6. The gamma group goes on to include Bangkok, Peking, Montreal, Rome, Stockholm and Warsaw, followed just below by Atlanta, Barcelona, Berlin, Budapest, Buenos Aires, Copenhagen, Hamburg, Istanbul, Kuala Lumpur, Manila, Miami, Minneapolis, Munich and Shanghai. As for the large group of cities which only have some indices of potential world status, with a score less than 4, their list can be found in the Bulletin (page citylist.html) on the site indicated by note 1.


see : world cities n° 2